Existing approaches to wireless networking, such as the example shown in network 100 of FIG. 1, use specialized hardware for user plane gateways and control plane signaling implementation. User traffic typically is routed on a long round-trip to the home network as shown in route 102, which makes it difficult to access local services in the visited network. In this example, because the control plane is handled by servers such as mobile management entity 114 and application server 118, along with serving gateway (SGW) 112 and packet data network (PDN) gateway (PGW) 116, data traffic is routed from base station 106 through routers 108 and 110 to SGW 112 to PGW 116 through PGW 120 and high bandwidth WAN to a Wireless Access Gateway (WAG) 121 and then through router 122 and base station 111 to its destination: user equipment 105. A more direct route would be to route the traffic from router 108 directly through router 122. This would apply to traffic through base stations 107 and 109 as well. However, this is not feasible with prior techniques because the routing decision occurs at a central location in a specialized wireless gateway. This increases capital expenditures due to amortization of development costs over fewer units because software and hardware designs are applicable only to large carrier-hosted wireless packet networks. This also increases the operating expenses of the operator due to the need to maintain high bandwidth wide-area network links to remote networks. It also makes it difficult to support applications that use knowledge of the local network such as the venue or enterprise in which the wireless base stations are deployed. Only operator applications are supported, and they must be processed through a policy and charging rules function (PCRF) to modify policies for the user plane. Further, enforcement happens at a remote PGW.